Multi-Agent Web App (mawa)

Project Overview

mawa (Multi-Agent Web App) is an experimental project designed to demonstrate and teach how to build a simple web application using the Google Agent Development Kit (ADK). This repository serves primarily as an educational resource, offering a hands-on approach to understanding agent-based systems and their integration into web environments with Google's tools.

While not intended for production use, mawa provides a practical learning environment. Learning new concepts by applying them to familiar problems is a powerful method, and this project aims to facilitate that process for those interested in the Google ADK and multi-agent architectures.

Getting Started

To run mawa locally, follow the instructions below.

Prerequisites

Before you begin, ensure you have the following installed:

• Python 3.9+: This project uses modern Python features.
• Poetry: A dependency management and packaging tool for Python. If you don't have it, you can install it by following the instructions on the Poetry website.

Installation

1. Clone the repository:

   git clone [https://github.com/your-username/mawa.git](https://github.com/your-username/mawa.git)
   cd mawa

2. Install dependencies using Poetry:

   poetry install

Configuration

To enable the application to interact with the Google Gemini API, you need to set up your API key:

1. Navigate to the src/mawa directory:

   cd src/mawa

2. Create a new file named .env in this directory.

3. Add the following content to the .env file, replacing <your gemini API key> with your actual Google Gemini API key:

   GOOGLE_GENAI_USE_VERTEXAI=FALSE
   GOOGLE_API_KEY=<your gemini API key>

   Note: You can obtain a Google Gemini API key from the Google AI Studio.

Running the Application

You have two primary methods to run the mawa application: using the ADK web server or directly via PyCharm.

1. Running with ADK Web

This method utilizes the Google ADK's built-in web server capabilities.

1. Ensure you are in the src directory:

   cd src

   (If you are still in src/mawa, navigate back one level: cd ..)

2. Execute the following command to start the ADK web server:

   poetry run adk web

   The application should now be accessible in your web browser, typically at http://localhost:8000.

2. Running in PyCharm

For developers using PyCharm, you can configure a run configuration to launch the application.

1. Open your project in PyCharm.

2. Create a new Python Run/Debug Configuration:

   • Go to Run -> Edit Configurations...
   • Click the + icon and select Python.

3. Configure the settings:

   • Module name: Select uvicorn
   • Parameters: Set to mawa.main:app --reload --app-dir src
   • Environment variables: Add GOOGLE_API_KEY=<your gemini API key> (replace with your actual key). You can click the folder icon to easily add this variable.

4. Apply the changes and Run the configuration.

---

Disclaimer

This project is purely for educational purposes and to demonstrate concepts related to the Google Agent Development Kit. It is not designed for production environments and may not follow best practices for security, scalability, or robustness required for such use cases. Use it as a learning tool to explore the capabilities of ADK and multi-agent web applications.